The Global Trade Analysis Project (GTAP) 6 Data Base aims to represent the world economy for the reference year 2001. GTAP 6 provides a set of files that contain flows of goods and services within and among regions based on regional Input-Output (I-O) Tables for 87 regions and 57 commodities. All terms are in money values, in millions of 2001 U.S. dollars. Except for energy, no volume data is supplied. The GTAP 6 Data Base is organized in 4 separate data packages which contain the file sets, parameters, main data and energy volume data.
Figure 1 shows the general structure of an I-O table. Each row describes the output of a particular industry (i.e., the value of the commodity it produces that is used by the other sectors of the economy) or, in the value added section of the table, the allocation of an endowment across sectors. K defines capital, L denotes labor and T includes any type of taxes. Each column describes the inputs to an industry (i.e., the value of the commodities or endowments it purchases as input to its production). Final demand includes private (C) and government (G) consumption, investment (I), exports (X) and imports (M). The total cost of production for a particular industry is represented by the column sums, and the total value of the output of a particular industry is represented by the row sums (including exports, minus imports).
GTAP data are based primarily on national I-O tables, but in the cases of energy, agriculture and food, trade, and import and export duties, values in the national tables are substituted by separately constructed data bases. This procedure is intended to ensure international consistency, particularly regarding the sectoral mapping of commodities, which is performed in different ways across countries in the original I-O tables.
C I G X M
K L T
Inputs to production Final Demand
Outputs of production
Value of intermediate usage of commodity j by producers of commodity i
Primary goods
Value added used by final demand
Inputs to industry jInputs to industry j
Total cost of producing commodity j National Accounts
Tot. val. add.Total output of commodity i Ouputs of industry i
Figure 1. General structure of an Input-Output table.
2.1 Intermediate production
Since the GTAP database is designed primarily to support studies involving international trade, intermediate and final demand are divided into purchases from
domestic sources and those purchased from abroad. Our interest is limited to the overall intermediate and final demand; for this reason we sum up imports and domestic purchases. Figure 2 shows a decomposition of a hypothetical intermediate production sector and the units in which data are supplied, in order to provide a basis for further discussion of the data.
Figure 2. Decomposition of inputs to intermediate production into domestic and imported purchases. Different price measures employed are shown in blue boxes.
All intermediate flows of goods and services are available in terms of agents’
and market prices. Agents’ values of commodities are the sum of the market value of the respective commodity and net taxes (taxes minus subsidies). This applies to all goods in the market, for intermediate production as well as for primary production and final demand. We use market prices and include net taxes as additional row in the I-O table.
2.2 Primary factors and taxes on inputs and outputs
In Figure 1, we display labor and capital as the only endowment commodities. But in fact, GTAP distinguishes 5 different primary factors, namely skilled and unskilled labor, land, natural resources and capital. GTAP separates labor into skilled and unskilled labor on the basis of occupational classifications. In addition to GTAP 6 there is a separate land use data base which splits up land into agro ecological zones (AEZ’s) and their use by the different industries.
GTAP provides figures on income tax, by factor and region (not by industry), as well as factor employment tax, by factor, industry and region, which is, again, the
difference between market value and agents’ value of output (Figure 2). In addition, there are taxes on outputs (ad valorem tax) by commodity (summed over industries).
We include the net value of all taxes (also export and import taxes) as a single row in the value added portion of the I-O table.
2.3 Final demand
For investment goods, which are based on changes in physical capital stocks, GTAP creates a fictitious sector. Inputs to this sector represent imported and domestic purchases of goods used for investment (e.g. machinery, construction of buildings, etc.).
Although GTAP provides data on investment for the economy as a whole broken down by commodities, it does not characterize investment undertaken by each sector.
Residential and government consumption, again, is split up into domestic and imported purchases, which can be easily summed up sector by sector.
2.4 Data aggregation
GTAP 6 is supported by GTAPAgg, a Windows program that aggregates several regions/commodities to one or more common regions/sectors and creates a data base of the selected collection of countries/commodities. Besides regions and sectors, primary factors can also be aggregated. GTAPAgg is very easy to use and provides an interface to specify the new (aggregate) regions/commodities you want to create out of the corresponding GTAP regions/commodities. Once the data is aggregated, it can be exported without any difficulties and manipulated by a spreadsheet program (e.g. MS Excel).
GTAPAgg aggregates countries by a simple addition of the elements of the I-O table from several countries to make a region. For example, in creating a European region, the value of energy inputs to Albania’s transportation sector is added to the corresponding value for Bulgaria. This aggregation introduces an inconsistency regarding imported and domestic production. In GTAP, inputs to each sector are disaggregated into domestic and imported purchases. Adding these separately, as is done in the GTAPAgg program, leads to an overestimate of imported purchases (and underestimate of domestic purchases) in the aggregate region, because some portion of purchases previously defined as imported (e.g. from Albania to Bulgaria) should now be defined as domestic production (i.e. within Europe).
To adjust for this inconsistency in the aggregated data, the intra-regional trade has to be subtracted from each industrial sector’s imported purchases and shifted to its domestic purchases. Because in the PET model we do not differentiate between domestic and imported inputs to production, we are not affected by this problem.
We apply the GTAP production data to the structure of the PET model. Our aim is to deploy up to 25 production sectors in the PET model (see Appendix Table 1) and up to 24 regions (see Appendix Table 2), but in the current version of the model we use 9 regions and 5 sectors: “coal”, “oil and gas”, “refined fuels”, “electricity” and
“materials”. In principle one could use any aggregation of the 57 GTAP production sectors. Figure 2 gives the structure of the aggregated I-O table we employ in the PET model. Note that in addition to aggregating industries, we also aggregate taxes into a
single net tax row, aggregate land, natural resources, and capital into a single capital row, and aggregate skilled and unskilled labor into a single labor row.
coal oil + gas ref. fuels electr. mater.
coal
Figure 3. Structure of the Input-Output table currently used in the PET model.